Executive Summary : | Recently, there has been a growing interest in designing the cluster trials using stepped wedge design (SWD). A SWD design is a type of cluster–crossover design in which clusters of individuals are randomized unidirectional from a control to an intervention at certain time points. SWDs are more applicable to implementation studies in which the staggered change from one treatment to another treatment can have practical advantages. Moreover, due to some logistical, practical or financial constraints the intervention can only be implemented in stages. Similar to the parallel CRTs, the ICC affects the SWDs. A commonly used standard SWD dictates an equal number of clusters (balanced design) be switched to treatment at each period. We have shown that an optimal SWD may be very sensitive to the unknown ICC (see Example 3.1). In this project we aim to achieve the following goals: 1. Proposing optimal cluster size w.r.t. the appropriate priors assigned to the ICC. 2. Extending the proposed methodology to the cohort SWDs. 3. Obtaining optimal SWDs for the non–normal response (for generalized linear models). |